Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

lookup: complement data from context with general text statistics #5479

Merged
merged 5 commits into from Mar 23, 2024
Merged

lookup: complement data from context with general text statistics #5479

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
a51a401
lookup: evaluation tools, use corpus/previous gens
JohannesGaessler Feb 11, 2024
5070d0a
fixup! lookup: evaluation tools, use corpus/previous gens
JohannesGaessler Mar 22, 2024
e2f6704
fixup! lookup: evaluation tools, use corpus/previous gens
JohannesGaessler Mar 22, 2024
e04cf1a
fixup! lookup: evaluation tools, use corpus/previous gens
JohannesGaessler Mar 22, 2024
63d03a9
fixup! lookup: evaluation tools, use corpus/previous gens
JohannesGaessler Mar 22, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
3 changes: 3 additions & 0 deletions .gitignore
View file
Open in desktop
Expand Up @@ -58,6 +58,9 @@ models-mnt
/llava-cli
/lookahead
/lookup
/lookup-create
/lookup-merge
/lookup-stats
/main
/metal
/passkey
Expand Down
13 changes: 11 additions & 2 deletions Makefile
View file
Open in desktop
Expand Up @@ -676,14 +676,17 @@ json-schema-to-grammar.o: common/json-schema-to-grammar.cpp common/json-schema-t
train.o: common/train.cpp common/train.h
$(CXX) $(CXXFLAGS) -c $< -o $@

ngram-cache.o: common/ngram-cache.cpp common/ngram-cache.h
$(CXX) $(CXXFLAGS) -c $< -o $@

libllama.so: llama.o ggml.o $(OBJS)
$(CXX) $(CXXFLAGS) -shared -fPIC -o $@ $^ $(LDFLAGS)

libllama.a: llama.o ggml.o $(OBJS) $(COMMON_DEPS)
ar rcs libllama.a llama.o ggml.o $(OBJS) $(COMMON_DEPS)

clean:
rm -vrf *.o tests/*.o *.so *.a *.dll benchmark-matmult common/build-info.cpp *.dot $(COV_TARGETS) $(BUILD_TARGETS) $(TEST_TARGETS)
rm -vrf *.o tests/*.o *.so *.a *.dll benchmark-matmult lookup-create lookup-merge lookup-stats common/build-info.cpp *.dot $(COV_TARGETS) $(BUILD_TARGETS) $(TEST_TARGETS)
find examples pocs -type f -name "*.o" -delete

#
Expand Down Expand Up @@ -813,9 +816,15 @@ lookahead: examples/lookahead/lookahead.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)

lookup: examples/lookup/lookup.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
lookup: examples/lookup/lookup.cpp ggml.o llama.o ngram-cache.o $(COMMON_DEPS) $(OBJS)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
$(CXX) $(CXXFLAGS) -c examples/lookup/lookup-create.cpp -o $(call GET_OBJ_FILE, examples/lookup/lookup-create.cpp)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, examples/lookup/lookup-create.cpp) -o lookup-create $(LDFLAGS)
$(CXX) $(CXXFLAGS) -c examples/lookup/lookup-merge.cpp -o $(call GET_OBJ_FILE, examples/lookup/lookup-merge.cpp)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, examples/lookup/lookup-merge.cpp) -o lookup-merge $(LDFLAGS)
$(CXX) $(CXXFLAGS) -c examples/lookup/lookup-stats.cpp -o $(call GET_OBJ_FILE, examples/lookup/lookup-stats.cpp)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, examples/lookup/lookup-stats.cpp) -o lookup-stats $(LDFLAGS)

passkey: examples/passkey/passkey.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
Expand Down
2 changes: 2 additions & 0 deletions common/CMakeLists.txt
View file
Open in desktop
Expand Up @@ -65,6 +65,8 @@ add_library(${TARGET} STATIC
json.hpp
train.h
train.cpp
ngram-cache.h
ngram-cache.cpp
)

if (BUILD_SHARED_LIBS)
Expand Down
20 changes: 20 additions & 0 deletions common/common.cpp
View file
Open in desktop
Expand Up @@ -963,6 +963,22 @@ static bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg,
}
return true;
}
if (arg == "-lcs" || arg == "--lookup-cache-static") {
if (++i >= argc) {
invalid_param = true;
return true;
}
params.lookup_cache_static = argv[i];
return true;
}
if (arg == "-lcd" || arg == "--lookup-cache-dynamic") {
if (++i >= argc) {
invalid_param = true;
return true;
}
params.lookup_cache_dynamic = argv[i];
return true;
}
if (arg == "--save-all-logits" || arg == "--kl-divergence-base") {
if (++i >= argc) {
invalid_param = true;
Expand Down Expand Up @@ -1436,6 +1452,10 @@ void gpt_print_usage(int /*argc*/, char ** argv, const gpt_params & params) {
printf(" Hugging Face model file (default: unused)\n");
printf(" -ld LOGDIR, --logdir LOGDIR\n");
printf(" path under which to save YAML logs (no logging if unset)\n");
printf(" -lcs FNAME, --lookup-cache-static FNAME\n");
printf(" path to static lookup cache to use for lookup decoding (not updated by generation)\n");
printf(" -lcd FNAME, --lookup-cache-dynamic FNAME\n");
printf(" path to dynamic lookup cache to use for lookup decoding (updated by generation)\n");
printf(" --override-kv KEY=TYPE:VALUE\n");
printf(" advanced option to override model metadata by key. may be specified multiple times.\n");
printf(" types: int, float, bool. example: --override-kv tokenizer.ggml.add_bos_token=bool:false\n");
Expand Down
28 changes: 15 additions & 13 deletions common/common.h
View file
Open in desktop
Expand Up @@ -88,20 +88,22 @@ struct gpt_params {
// // sampling parameters
struct llama_sampling_params sparams;

std::string model = "models/7B/ggml-model-f16.gguf"; // model path
std::string model_draft = ""; // draft model for speculative decoding
std::string model_alias = "unknown"; // model alias
std::string model_url = ""; // model url to download
std::string hf_repo = ""; // HF repo
std::string hf_file = ""; // HF file
std::string prompt = "";
std::string prompt_file = ""; // store the external prompt file name
std::string path_prompt_cache = ""; // path to file for saving/loading prompt eval state
std::string input_prefix = ""; // string to prefix user inputs with
std::string input_suffix = ""; // string to suffix user inputs with
std::string model = "models/7B/ggml-model-f16.gguf"; // model path
std::string model_draft = ""; // draft model for speculative decoding
std::string model_alias = "unknown"; // model alias
std::string model_url = ""; // model url to download
std::string hf_repo = ""; // HF repo
std::string hf_file = ""; // HF file
std::string prompt = "";
std::string prompt_file = ""; // store the external prompt file name
std::string path_prompt_cache = ""; // path to file for saving/loading prompt eval state
std::string input_prefix = ""; // string to prefix user inputs with
std::string input_suffix = ""; // string to suffix user inputs with
std::vector<std::string> antiprompt; // string upon seeing which more user input is prompted
std::string logdir = ""; // directory in which to save YAML log files
std::string logits_file = ""; // file for saving *all* logits
std::string logdir = ""; // directory in which to save YAML log files
std::string lookup_cache_static = ""; // path of static ngram cache file for lookup decoding
std::string lookup_cache_dynamic = ""; // path of dynamic ngram cache file for lookup decoding
std::string logits_file = ""; // file for saving *all* logits

std::vector<llama_model_kv_override> kv_overrides;

Expand Down